Python Implementation of Point Cloud Reconstruction from Two Cameras

Sign in to Add to Watch List

The client has made the following changes to the job.

Description

Date

Job awarded.

Apr 18, 2014

Bidding period extended.

Apr 10, 2014

Skills updated.

Apr 5, 2014

Attachment added: TwoViewTriangulationBrief.pdf

Attachment removed: TwoViewTriangulationBrief.pdf

Apr 3, 2014

Job description changed. New Version|Previous Version

Attachment added: TwoViewTriangulationBrief.pdf

Attachment removed: TwoViewTriangulationBrief.pdf

Apr 3, 2014

Updated

Overview Neapolitan is interested in determining the structure of a 3 dimensional object using 2 cameras. Our approach is to model the object as a point cloud. The goal of this job is to compute the 3 dimensional coordinates of each point via triangulation using images of the points in each camera. Detailed Description The assumption is that in a previous image processing step, the pixel locations of the image of each point have already been computed. Each camera is also known both intrinsically and extrinsically. In other words, the cameras have been calibrated and the K matrix is available. Also, the position and orientation of each camera is known (R, t). These are the inputs to the program. The program should compute pixel correspondences between images and triangulate the location of each point. Since there is no scene information to compute correspondences, a guess and check method may need to be implemented. Best practices should be used to optimize this process, avoiding the implementation of a naive brute force method – improving computation time. For example, potential correspondence candidates should be chosen using epipolar geometry. It should not be assumed that both cameras see all points. In general there will be n pixel locations associated with image 1 and m pixel locations associated with image 2, where n not equal to m. Exact point triangulations should also not be expected due to errors and noise in the pixel location computation. Successful (and unsuccessful) triangulations should be determined by examination of the triangulation residuals. A .pdf of this description as well as the deliverables are attached below.

Overview The purpose of this job is to determine the structure of a 3 dimensional object using 2 cameras. The approach is to model the object as a point cloud. The goal of this job is to compute the 3 dimensional coordinates of each point via triangulation using images of the points in each camera. Detailed Description The assumption is that in a previous image processing step, the pixel locations of the image of each point have already been computed. Each camera is also known both intrinsically and extrinsically. In other words, the cameras have been calibrated and the K matrix is available. Also, the position and orientation of each camera is known (R, t). These are the inputs to the program. The program should compute pixel correspondences between images and triangulate the location of each point. Since there is no scene information to compute correspondences, a guess and check method may need to be implemented. Best practices should be used to optimize this process, avoidin...

Apr 3, 2014

**Client prefers freelancers from:Anywhere**

You are still able to submit a proposal for this job.

**The client prefers freelancers from a different location.**

Overview

The purpose of this job is to determine the structure of a 3 dimensional object using 2 cameras. The approach is to model the object as a point cloud. The goal of this job is to compute the 3 dimensional coordinates of each point via triangulation using images of the points in each camera.

Detailed Description

The assumption is that in a previous image processing step, the pixel locations of the image of each point have already been computed. Each camera is also known both intrinsically and extrinsically. In other words, the cameras have been calibrated and the K matrix is available. Also, the position and orientation of each camera is known (R, t). These are the inputs to the program.

The program should compute pixel correspondences between images and triangulate the location of each point. Since there is no scene information to compute correspondences, a guess and check method may need to be implemented. Best practices should be used to optimize this process, avoidin...

Sign in or Register to see more

Job ID: 55183953

Avg $ | High $ | Low $ — Show Pricing

- Submit Date (Latest)

Bid ID: 55609034 | Submitted: Apr 14, 2014 05:34 ET

Delivery

Within 1 week

Within 1 week

$320.00

India

INVITED

A computers science professional specialized in the Image Processing, Pattern Recognition, and Machine Learning. I've worked on several projects on...

Bid ID: 55498379 | Submitted: Apr 10, 2014 23:39 ET

Proposal SEALED

Muhammad Awais Akhter

Muhammad Awais Akhter's identity has been verified.
Learn More

Belgium |

I am PhD student in area of remote sensing and image processing. Did my MPhil from UK in image processing. I have over 10 years job experience in...

Bid ID: 55260174 | Submitted: Apr 5, 2014 16:21 ET

Proposal SEALED

Belgium |

Doing projects in DSP, Image Processing, Computer Vision, Pattern Recognition, mathematical algorithm implementation etc.
Working since 2005 on...

Bid ID: 55256519 | Submitted: Apr 5, 2014 13:38 ET

Proposal SEALED

Elance is now an Upwork company.

Upwork is the choice of 4M+ clients. Get started working on Upwork today.

Are you ready to post a job like this one?

Post a Similar Job »